Uncooled Astro Cameras: Should You Consider Them And How To Use Them

The question is simple: should you consider getting an uncooled dedicated astro camera such as the ZWO ASI183MC or ASI294MC? The answer, though, is not that simple.

When asking such a question on the internet, statistically you will get a bunch of “don’t bother”, “just buy a cooled camera”, “just stick to your DSLR and save for a cooled camera”, and Co.

It is clear that uncooled astro cameras don’t get much love these days. But are they so worthless as anyone seems to think?

Since I have an uncooled ZWO ASI183MC, I decided to share with you my experience in the most objective way possible.

Samples From My Deep Sky Astrophotography With The Uncooled ASI183MC

Before going technical, here is a sample of my work with my uncooled ASI183MC (click on the image to enlarge them).

And here is the Bubble Nebula. It was taken with an even more improbable camera for deep sky astrophotography: the ZWO ASI224MC.

The Bubble Nebula, with ASI224MC fitted with the Optolong L-Enhance filter on Sky-Watcher Evoguide 50ED. Tracked with Sky-Watcher Az-GTI, unguided. 5hrs total integration time with 120” subs.

This camera has an even smaller sensor than the 1”-type 20MP sensor of the ASI183MC, but has no UV/IR cut filter, extremely low read noise, and high quantum efficiency.

Although the ASI224MC clearly shines in lunar and planetary astrophotography, it can occasionally be used to play the game of deep-sky astrophotography.

Sold already? No? Ok, let’s dig deeper then :)

Cooled Vs Uncooled: Principles

Uncooled cameras have no active way to get rid of the heat generated during the recording of the image. There are little vents on the side of the camera body, but this is about it. Not even sure if the body acts as a heat skin.

With my ASI183MC, my sensor temperature typically is anywhere between 5 to 10ºC above ambient temperature.

Cooled cameras, on the other hand, use a combination of thermoelectric cooling (Peltier) and a fan to actively dissipate the heat generated by the sensor. The cooled version of the camera I have allows dropping the sensor temperature to a max of 40-45ºC below ambient temperature.

The Problem With The Sensor Temperature

The main problem of not having a cooled sensor is noise. Typically read noise, the noise generated when your camera reads the charge of a pixel, double every 5ºC of increase in temperature. Thermal noise and the number of hot pixels also increase with temperature.

The image below compares two crops 100% for a single light frame taken with the same gain and exposure, but different temperatures.

Noise comparison between single 300” sub taken at a temperature of 21ºC (top) and 1ºC (bottom).

As you can see, the lower the sensor temperature, the less noisy your image will be.

But the amount of noise is not the only problem: with uncooled cameras, you cannot keep stable the sensor temperature, which will change accordingly to the ambient temperature. This makes it difficult to match dark frames to the lights. More on this later.

The plot below shows the drifting of my AS183MC sensor temperature during a 3-hour session (36x300” images) I did at the beginning of November.

The drift of my sensor temperature during 3 hrs long session with 300” subs. The average temperature is 14ºC with a standard deviation of 0.5ºC. Data from the fits.

As you can see, the temperature dropped during the night, although only by less than 2ºC.

Uncooled Vs Cooled Astro Cameras

Here, of course, the cooled camera is the winner, particularly if you have high-end astrophotography gear. Who would like to spend tons of money on a quintuplet refractor on an EQR-6 PRO mount, guiding system, and all the whistles and bells and cripple the image quality with thermal and read noises?

On the other hand, if one is a beginner or amateur astrophotographer with photographic lenses or entry-level scopes on a small mount, uncooled cameras may be attractive. The 2-300$ difference with respect to getting the cooled version can be invested in filters or other needed accessories.

Cooled cameras also need to have a 12V DC power source and desiccant tabs to prevent fogging when cooling the sensor below ambient temperature.

One thing to consider is that you can easily re-sell a cooled camera, while uncooled ones may be more difficult to sell, should you decide to upgrade (or give up).

DSLR Vs Uncooled Astro Cameras For Deep Sky Astrophotography

Noise is the only real difference between a cooled and uncooled camera, and image quality is objectively superior with cooled cameras.

But how bad actually is an uncooled camera for beginners and amateur astrophotographers like us? Are people saying “for deep sky astrophotography you may well keep shooting with your DSLR rather than getting an uncooled camera, particularly if it has small pixels” right?

I believe they are not and here are my top 5 arguments for believing so:

Uncooled cameras are smaller and lighter than DSLRs, helping reduce the payload and improving balance if you are on a tracker or lightweight equatorial mount. They are also easier on entry-level instruments with small/weak focuser;
Uncooled cameras are great for deep sky astrophotography as they do not have the UV/IR cut filter in front of the sensor DSLRs have. This allows you to begin taking advantage of narrowband filters for color cameras, such as the Optolong L-Enhance.
Particularly if you don’t have a top-of-the-line DSLR, any uncooled camera has lower read noise. My ASI183MC has a read noise ranging from 3 to 1.6, depending on the gain used. At unity gain, it has a read noise of 2.3. Depending on the ISO used, PhotontoPhoto reports for the Canon 60D a read noise ranging from 2.6 to 6.1;
Quantum Efficiency, the ability to convert photons to electrons, is usually much higher with dedicated astro cameras. The higher the QE, the more sensitive the sensor is (not to be confused with the concept of gain or ISO, which is a mere digital amplification of the signal coming from the sensor). My ASI183MC has a QE of 84% (84% of the incoming photons are converted to an electrical signal), while the Canon 60D has, in comparison, a puny 40% QE.
Astro cameras usually have higher compatibility with accessories for astrophotography and in particular for filters. Some clip-in filters are available for some DSLR and mirrorless, but not for all.

Read Noise Vs ISO for the Canon 60D. Image credit: PhotonstoPhoto.

QE for the Canon 60D. Image Credit: PhotononPhoto.

Converting your DSLR to Ha or full-spectrum will cost anywhere from 200 to 500USD, depending on the camera and the lab chosen to do that professionally. Plus shipping. Plus UV/IR cut filter for normal daylight photography.

And to cut costs, people often suggest getting an older DSLR and having it converted. With the cost of the camera, the conversion, batteries, and accessories you may get close to the cost of an uncooled astro camera and you have an older sensor, probably noise anyway. Is it worth it? I doubt.

The Light Frames Calibration Problem

You get what you pay for.

Everybody seems to love this saying and often this is used in a negative way: “oh yes is a noisy image. You get what you pay for”.

What everybody seems to forget these days is that if it is true that you get what you pay for, it is also true you obtain what you know about. With this, I mean that a lot can be done with relatively budget equipment, if one is willing to learn the most about it and to be creative, rather than just pay for a better piece of equipment.

We already said the big problem with using an uncooled camera is noise.

Noise is reduced with taking darks, calibration frames were taken with the same exposure time, gain/ISO and with the lens/telescope cap on, so that light will not be reaching the sensor.

In short, darks account for read noise and thermal noise. And because they deal with this noise, they are temperature-dependent and you should find a way to match the temperature of your darks with that of your light frames at the best of your possibilities.

A cooled camera is great: since the sensor temperature is under your control and mostly independent from the ambient temperature, you can take your darks whenever you want, thus using all your night time in location to actually photograph the sky, rather than wasting time with getting darks.

If you have an uncooled camera, they say, you have to take dark while being in a location, and this cuts down the time you can photograph the sky. Particularly if you are into long subs: to take 20 dark (a good number for calibrating your lights) and you shot 300” light frames, you will be wasting almost 2hrs collecting no light at all.

If you have DSLR cameras, you may get away with no darks at all if you dither between shots and if you take bias frames, or you can use masterwork scaling to let the calibration software adjust your darks so as to match the temperature, exposure, and gain/ISO used for your light frames.

Dealing With Amp Glow In Uncooled Cameras

People bashing uncooled cameras often mention amp glow as a reason to do so. Some camera models, including cooled cameras, suffer from amp glow, visible as streaks of light creeping into the frame.

The Amp Glow from my ASI183MC is clearly visible in this dark frame.

The image above clearly shows the amp glow in a 300” dark frame taken with may uncooled ASI183MC and gain 149. In the image below, you see the amp glow creeping into a single light frame. In both cases, the sensor temperature was about 21ºC.

People then argue this amp glow is hard to calibrate out if you can’t match the darks temperature to that of the light frames. At least DSLR cameras don’t show amp glow (or not a strong one anyway).

Well, how hard it is to calibrate out this amp glow in real life? As hard as collecting 16 darks.

Below you see the calibrated, stacked image with 16 darks. Note that I have used the strongest automatic stretch Astro PIxel Processor offers to try making any residual amp glow visible in the image.

The stacked and calibrated image only show very little residual amp glow with such aggressive stretching.

The good news is that there is very little residual amp glow, visible only with such aggressive image stretching, something you will not normally do.

And how well does the temperature for the darks match that for the light frames?

Comparison between the sensor temperature for the light frames and dark frames. Data from the fits files.

As you can see from the plot above, while the average temperature for the light frames and dark frames are quite close, there are large fluctuations up to 3ºC. Yet, the amp glow calibrates out real good.

So, how close does the temperature of the darks have to match that of the light frames? Ideally, you want to have a perfect match for the best calibration possible. In practice, differences smaller than 5ºC are acceptable, in my experience.

And this opens up the possibility to adopt strategies for getting your darks.

“Spray & Pray”: A Viable Strategy To Collect Darks For Your Uncooled Camera

When you have amp glow to remove, darks cannot be adapted to the light frames using scaling algorithms and you need to get them at same exposure time, gain, and, possibly, the temperature of images you want to calibrate.

In photography, “Spray&Pray” refers to the action of taking a lot of photos in the hope that few will be good. Here, I take darks every now and then, during the day or cloudy night, in the hope to have covered all the most common ambient temperatures I will be imaging with.

Darks taken in the middle of the day during winter, for example, could match quite well the temperature of light frames taken at night during spring or even summer.

By keeping an eye on the temperature of your light during particular periods of the year you can get a good idea of when collecting some darks during the day or cloudy nights.

And since exposure time mostly depends on the sky's brightness and the use of filters, with time and practice, you will find yourself often using the same exposure and gain settings for your photography. For example, I mostly use a gain of 149 and 300” exposure time when working with the Optolong L-Enhance filter.

At the end of the day, you will have only a few combinations where the long exposure makes getting darks in location a pain in the neck.

If I take images shorter than 1 minute or 90s, I get all the darks I need while packing.

The Spray & Pray Method In Real Life

Let’s consider this session on the Flaming Star Nebula in Auriga.

Flaming Star Nebula in Auriga. 4hrs total integration time, gain 149 with 300” subs.

The average temperature of the light frames, extracted from the fits, was 14ºC. I asked my script to extract all the darks in the library with temperature within 1ºC and 2ºC from the average temperature of the lights.

With the first criteria, the script fetched 16 darks while with the second criteria it fetched 60 darks.

Temperature evolution for the lights and the two sets of darks fetched by my script.

It is worth nothing that the 60 darks have been collected in 5 different nights around the night I was out imaging.

Below is the comparison of a 100% crop for the residual noise in the calibrated (and aggressively stretched) images with the two sets of darks: despite the larger difference in temperature, the set with 60 darks cleaned up the image better. Taking 60 darks in the field would have not been possible, as the process takes 5hrs.

On the left the calibrated image with 16 darks within 1ºC from the average temperature of the light frames. On the right the calibrated image with 64 darks within 2ºC from the average temperature of the light frames.

Note that all the images you see here have not being dithered (I know, right?) because dithering takes time, particularly if you do it for every image you take. But dithering is something that is be greatly beneficial to your image quality so, with the need of taking darks in the field out of the way, you can now investa part of the extra imaging time you got in dithering.

Scripts For Pray&Spray

If you have an uncooled camera and want to try my method, I made three scripts to:

rename the files with their creation date, so that from time to time I can remove the oldest one in my library, and no files are overwritten;
get the temperature for the light frames I need to calibrate;
extract from the dark library the dark frames that are in the desired range of temperature around the average temperature of the light frames.

They are bash scripts and I tested them on my Mac. Of course, there is no warranty, so before start playing make sure you have a backup of your data. I take no responsibility.

ap_rendarks.sh (Download)

This script simply renames your darks with their date and time of creation.

Usage: simply type ./ap_rendarks.sh in the directory of your darks.

After you have renamed the files, move them into the directory containing your dark library.

ap_gettemp.sh (Download)

This script uses the fitsheader utility to extract gain, exposure time, and temperature for all fit files in the current directory.

Usage: ./ap_gettemp.sh to analyze fit files in the current directory

The script also calculates the average temperature and its standard deviation. The name of each fit file, with gain, exposure time, and temperature, as well as the average temperature for the sequence and standard deviation are saved in a log file.
The script will warn you if you are mixing files with different exposure times and/or gain.

I use this script to get the temperature of my light frames.

ap_darksfinder.sh (Download)

This script should stay in the directory with all your darks and it fetches those that satisfy your criteria about exposure time, gain, and temperature range. You can specify the max number of darks to fetch or ask to get all the darks that fit the criteria (up to 1000… you can change this in the script).

Usage: ./ap_darksfinder -e exposure (in seconds) -g gain -t temperature -r temperature range -n number darks or ‘all’

e.g. ./ap_darksfinder -e 300 -g 149 -t 14 -r 2 -n all

will fetch all available darks shot for 300s at gain 149 whose temperature is 14±2ºC. All selected darks will be copied into a new directory with an evocative name. Running the script multiple times with the same parameters creates multiple directories.

Finally, the script will create a log, with the name, exposure time, gain and temperature of the fetched files, average temperature, and other info.

Conclusions

Uncooled cameras are, to me, superior to most entry-level or aging DSLR, even when modified. While they cannot compete with the cooled cameras for image quality, they are far to be the poor performers many people seem to think they are.

With a bit of tinkering and testing, it is perfectly possible to do decent astrophotography with uncooled cameras, particularly if one has entry-level/budget equipment.

If you can stretch your budget, go for a cooled camera. But if you cannot/want to and you are considering getting an older DSLR and having converted for astrophotography, I would suggest you consider a dedicated, uncooled, astro camera.